pooled_04_cytokine_repeats.Rmd
dbg_csa
5/4/2022
Loading
Sample Quality
ggplot(cyto.read.counts.combined) +
geom_col(aes(x=reorder(CAR.align, counts), fill=run, y=log10(counts))) +
facet_wrap(~sample.num+interaction(bin,sort.group), ncol=4) +
geom_hline(aes(yintercept=1), linetype=2) +
scale_x_discrete(guide = guide_axis(angle = 90)) ## Warning: Removed 565 rows containing missing values (geom_col).
mapped_only_plot <- ggplot(
cyto.read.counts.combined[,
list(mean_log10_reads= mean(counts, na.rm=T)),
by=c('sort.group','bin','sample.num','run')]) +
geom_col(
aes(
x=reorder(interaction(sort.group, bin, sample.num), mean_log10_reads),
y=mean_log10_reads, fill=run)) +
scale_x_discrete(guide = guide_axis(angle = 90)) +
scale_y_log10() +
labs(x='log-Mean Reads per CAR, by Sample', y='log-Mean Read Counts')
mapped_totals_plot <- ggplot(
cyto.read.counts.orig[,
CAR.masked := (CAR.align %in% c('41BB.wt','CD28.wt','*'))][,
list(count.sum= sum(counts.miniseq, na.rm=T)),
by=.(sort.group, bin, sample.num,CAR.masked)]) +
geom_col(
aes(
x=reorder(interaction(sort.group, bin, sample.num), count.sum),
y=count.sum, fill=!CAR.masked)) +
scale_x_discrete(guide = guide_axis(angle = 90)) +
labs(x='Sample, ordered by total reads', y='Total Reads', fill='Reads Mapped?')
mapped_totals_plot <- ggplot(
cyto.read.counts.orig[,
CAR.masked := factor(
((CAR.align=='*')+ 2*(CAR.align %in% c('41BB.wt','CD28.wt'))),
labels=c('Mapped','Unmapped','Contaminant'))][,
list(count.sum= sum(counts.miniseq, na.rm=T)),
by=.(sort.group, bin, sample.num,CAR.masked)]) +
geom_col(
aes(
x=reorder(interaction(sort.group, bin, sample.num), count.sum),
y=count.sum, fill=CAR.masked), position='stack') +
scale_x_discrete(guide = guide_axis(angle = 90)) +
labs(x='Sample, ordered by total reads', y='Total Reads', fill='Reads Mapped?')
# pct masked vs cell count
cyto.read.pct.masked <- unique(
cyto.read.counts.orig[!is.na(counts.miniseq)][,
CAR.masked := factor(
((CAR.align=='*')+ 2*(CAR.align %in% c('41BB.wt','CD28.wt'))),
labels=c('Mapped','Unmapped','Contaminant'))][,
list(
cell.count, count.sum= sum(counts.miniseq, na.rm=T),
pct.masked= sum(counts.miniseq[CAR.masked!='Mapped'])/sum(counts.miniseq),
pct.contam= sum(counts.miniseq[CAR.masked=='Contaminant'])/sum(counts.miniseq),
pct.unmapped= sum(counts.miniseq[CAR.masked=='Unmapped'])/sum(counts.miniseq)),
by=.(sort.group, bin, sample.num)])
mapped_pct_plot <- ggplot(cyto.read.pct.masked) +
geom_label(aes(x=pct.masked, y=cell.count, label=sample.num)) +
scale_y_log10() +
scale_x_continuous(label=scales::percent) +
labs(
x='Percent of Unmapped/Contaminant Reads',
y='log10 Cell Count of Sample')
contam_pct_plot <- ggplot(cyto.read.pct.masked) +
geom_label(aes(x=pct.contam, y=cell.count, label=sample.num), color='blue') +
scale_y_log10() +
scale_x_continuous(label=scales::percent) +
labs(
x='Percent of Contaminant Reads',
y='log10 Cell Count of Sample')
unmapped_pct_plot <- ggplot(cyto.read.pct.masked) +
geom_label(aes(x=pct.unmapped, y=cell.count, label=sample.num), color='green') +
scale_y_log10() +
scale_x_continuous(label=scales::percent) +
labs(
x='Percent of Unmapped Reads',
y='log10 Cell Count of Sample')
mapped_totals_plot /
(mapped_pct_plot | contam_pct_plot | unmapped_pct_plot)
Stats for Repooling 03.22.22
repool_df <- cyto.read.counts.orig[!is.na(counts.miniseq)][,
CAR.masked := factor(
((CAR.align=='*')+ 2*(CAR.align %in% c('41BB.wt','CD28.wt'))),
labels=c('Mapped','Unmapped','Contaminant'))][,
list(seq.sample= seq.sample[1], count.sum= sum(counts.miniseq), pcr2.ng.ul=pcr2.ng.ul[1]),
by=.(sort.group, bin, sample.num,CAR.masked)][CAR.masked == 'Mapped']
# rank counts lowest to highest
repool_df[, rank := rank(count.sum)]
# take bottom 23 (incl sample 7)
repool_num = 23
repool_df[rank <= repool_num, rel_pool_mass := sum(count.sum)/count.sum]
repool_df[rank <= repool_num, rel_pool_mass := rel_pool_mass/min(rel_pool_mass, na.rm=T)]
ggplot(repool_df) +
geom_col(
aes(
x=reorder(interaction(sort.group, bin, sample.num), count.sum),
y=count.sum, fill=CAR.masked), position='stack') +
scale_x_discrete(guide = guide_axis(angle = 90)) +
geom_vline(xintercept=repool_num+0.5, linetype=2) +
labs(x='Sample, ordered by total reads', y='Total Reads', fill='Reads Mapped?')
# set max mass ratio at 80x
repool_df[rank <= repool_num, rel_pool_mass := pmin(rel_pool_mass, 80, na.rm=T)]
# set absolute mass to 2 µL of lowest concentration sample
ul_sample_add <- 2
repool_df[rank <= repool_num, abs_pool_mass := min(pcr2.ng.ul) * rel_pool_mass *
ul_sample_add/(rel_pool_mass[which.min(pcr2.ng.ul)])]
# absolute volume of sample if undiluted
repool_df[rank <= repool_num, abs_pool_vol := abs_pool_mass/pcr2.ng.ul]
# volume of diluted sample, if applicable (if less than sample add vol)
repool_df[, vol_of_diluted := ifelse(abs_pool_vol > 2, abs_pool_vol, 2)]
# amount of h2o to add to diluted samples
repool_df[vol_of_diluted == 2, h2o_add :=
((ul_sample_add*pcr2.ng.ul)/(abs_pool_mass/2) - ul_sample_add)]
# check final masses
repool_df[, final_conc := ifelse(!is.na(h2o_add),
(pcr2.ng.ul * ul_sample_add) / (ul_sample_add + h2o_add),
pcr2.ng.ul)]
repool_df[, final_mass := final_conc * vol_of_diluted]
repool_df[, check := final_mass == abs_pool_mass]
fwrite(repool_df[rank <= 23],
'/Volumes/HFSE/Box/tcsl/ngs_data/2022.03.07.illumina_15/uniq/repool_table.csv')
repooled.samples <- repool_df[rank <= 23]$sample.numIntegrate nova and mini data
The rarest cars per sample appear to be higher in frequency in the novaseq - (more points in the upper left triangle). This is true across almost all samples, and esepcially non-repooled ones. Let’s compare the sort data between nova and mini for samples where most/all CARs are present for both.
ggplot(
cyto.read.counts.combined[counts > 0, sum(run=='nova') - sum(run=='mini'),
by=c('sample.num')]) +
geom_col(aes(x=reorder(sample.num,V1), y=V1, fill=factor(sign(V1)))) +
labs(x='Samples', y='Nova-Mini',
title='Red: More CARs in Mini, Blue: More CARs in Nova',
subtitle='Compare for samples in middle') +
geom_hline(yintercept=c(-6,6), linetype=2)
# add 3-way flag called run_choice: 'mini','nova','both'
cyto.read.counts.combined[,
run_choice := factor(ifelse(
abs(sum(run=='nova' & counts > 0) - sum(run=='mini' & counts > 0)) <= 6,
0,
sign(sum(run=='nova' & counts > 0) - sum(run=='mini' & counts > 0))),
levels=c(-1,0,1),
labels=c('mini','both','nova')),
by=c('sample.num')]Calc Bin Pcts
# replace miniseq data for repooled samples only
# cyto.read.counts <- rbind(
# cyto.read.counts.combined[sample.num %in% repooled.samples & run == 'nova'],
# cyto.read.counts.combined[!(sample.num %in% repooled.samples) & run == 'mini'])
calc_bin_stats <- function(cyto.read.counts) {
cyto.read.counts[,
bin.pct := cell.count/sum(cell.count),
by = .(CAR.align, sort.group)]
# get bin percent estimates
cyto.read.counts[,
bin.reads := sum(counts, na.rm=T),
by=c('sample.num')]
cyto.read.counts[,
car.bin.read.pct := counts/bin.reads,
by=c('sort.group','bin')]
# car bin read pcts should sum to 1
cyto.read.counts[bin.reads > 0,
stopifnot(all.equal(sum(car.bin.read.pct), 1, tolerance= 1e-5)),
by=c('sort.group','bin')]
# bin pcts should sum to 1
cyto.read.counts[, stopifnot(all.equal(sum(bin.pct), 1)),
by=c('sort.group','CAR.align')]
# get car absolute pct estimates
cyto.read.counts[,
car.abs.pct := car.bin.read.pct*(bin.pct),
by=c('sort.group')]
# absolute pct estimates should sum to 1
#cyto.read.counts[, (sum(car.abs.pct), 1)-1, by=c('sort.group')]
#very slightly off, 1/1000, need to fix tolerance
# get car abund
cyto.read.counts[, car.abund := sum(car.abs.pct),
by=c('sort.group','CAR.align')]
# CAR abundances should sum to 1
cyto.read.counts[,
stopifnot(all.equal(sum(car.abund), 1)),
by=c('sort.group','bin')]
cyto.read.counts[, car.bin.pct := car.abs.pct/car.abund]
# other metrics
cyto.read.counts[, rel.car.pct := (bin.pct - mean(bin.pct)) / mean(bin.pct),
by=c('sort.group','CAR.align')]
cyto.read.counts[, sort.reads := sum(counts),
by=c('sort.group','CAR.align')]
cyto.read.counts[, mean.car.abund := mean(car.abund),
by=c('CAR.align')]
cyto.read.counts[, rel.car.abund := car.abund/mean(car.abund),
by=c('CAR.align')]
cyto.read.counts[, car.bin.cells := car.bin.read.pct * cell.count]
cyto.read.counts[, car.sort.group.cells := sum(car.bin.cells),
by=.(CAR.align, sort.group)]
return(cyto.read.counts)
}
add_dpos_incl <- function(cyto.read.counts) {
# include dpos in ifng and il2
#il2
cyto.read.counts[, car.bin.pct.incl := car.bin.pct, by=.(sort.group, CAR.align)]
cyto.read.counts[, car.bin.pct.incl := ifelse(
bin=='il2',
car.bin.pct[bin=='il2']+car.bin.pct[bin=='dpos'],
car.bin.pct.incl),
by=.(sort.group, CAR.align)]
#ifng
cyto.read.counts[, car.bin.pct.incl := ifelse(
bin=='ifng',
car.bin.pct[bin=='ifng']+car.bin.pct[bin=='dpos'],
car.bin.pct.incl),
by=.(sort.group, CAR.align)]
return(cyto.read.counts)
}
min.sort.reads <- 200
cyto.read.counts.mini <- add_dpos_incl(calc_bin_stats(
cyto.read.counts.combined[
run_choice == run | (run_choice == 'both' & run == 'mini')]))
cyto.read.counts.nova <- add_dpos_incl(calc_bin_stats(
cyto.read.counts.combined[
run_choice == run | (run_choice == 'both' & run == 'nova')]))
# summary of read counts per sample per run
run.count.comparison <- merge(
cyto.read.counts.mini[,
list(run_choice=run_choice[1], mini.counts= sum(counts, na.rm=T)),
by=.(donor, sort.cond, sort.group)],
cyto.read.counts.nova[,
list(run_choice=run_choice[1], nova.counts= sum(counts, na.rm=T)),
by=.(donor, sort.cond, sort.group)])
ggplot(
melt(run.count.comparison, measure.vars=c('mini.counts','nova.counts'))) +
geom_col(aes(x=sort.cond, y=log10(value), fill=variable), position='dodge') +
facet_wrap(sort.group~interaction(donor, run_choice), scales='free')
# merge and replot bin pcts
run.count.comparison[, mini.greater := mini.counts < nova.counts]
cyto.read.counts.combined[, counts.sum := ifelse(
run_choice == 'both',
sum(counts, na.rm=T),
counts[run_choice == run]),
by=.(CAR.align, sample.num)]NovaSeq Repool Comparison (4.4.22)
ggplot(cyto.read.counts.orig[,
CAR.masked := factor(
((CAR.align=='*')+ 2*(CAR.align %in% c('41BB.wt','CD28.wt'))),
labels=c('Mapped','Unmapped','Contaminant'))][
CAR.masked == 'Mapped'][,
list(miniseq= sum(counts.miniseq*2, na.rm=T),
novaseq= sum(counts.novaseq*2, na.rm=T)),
by=sample.num],
aes(x=miniseq, y=novaseq, label=sample.num,
color=sample.num %in% repool_df[rank <= 23]$sample.num)) +
geom_point() +
scale_x_log10() +
scale_y_log10() +
geom_text_repel() +
geom_abline(linetype=2) +
labs(y='4.1.22 Novaseq Reads', x='3.16.22 Miniseq Reads',
color='Repooled for NovaSeq')## Warning: Transformation introduced infinite values in continuous y-axis
## Transformation introduced infinite values in continuous y-axis## Warning: ggrepel: 2 unlabeled data points (too many overlaps). Consider
## increasing max.overlaps
ggplot(
dcast(merge(run.count.comparison, cyto.read.counts.combined)[,
`:=`(log_counts= log10(counts), pct= counts / sum(counts)),
by=c('run','sample.num')][,
run_count_diff := log_counts[run=='nova'] -
log_counts[run=='mini'], by=c('CAR.align','sample.num')],
CAR.align + sample.num + run_count_diff + mini.greater ~ run , value.var='pct')) +
geom_point(
aes(x=mini, y=nova, color=sample.num %in% repooled.samples,
shape=factor(sign(run_count_diff)), size=abs(run_count_diff))) +
facet_wrap(~reorder(sample.num, sample.num %in% repooled.samples)) +
scale_x_log10() +
scale_y_log10() +
scale_size(range=c(0.5,3)) +
geom_abline() +
labs(x='Miniseq Read Fraction', y='Novaseq Read Fraction',
color='Repooled Sample', shape='sign(Novaseq - Miniseq) read count',
size='log10(abs(Novaseq - Miniseq)) read count')## Warning: Transformation introduced infinite values in continuous x-axis## Warning: Transformation introduced infinite values in continuous y-axis## Warning: Removed 56 rows containing missing values (geom_point).
# facet T/F - more reads in miniseq?
# for green (fewer than 6 car dropouts in each), combine reads without scaling.
ggplot(
dcast(merge(run.count.comparison, cyto.read.counts.combined)[,
`:=`(log_counts= log10(counts), pct= counts / sum(counts)),
by=c('run','sample.num')][,
run_count_diff := log_counts[run=='nova'] -
log_counts[run=='mini'], by=c('CAR.align','sample.num')],
CAR.align + sample.num + run_count_diff + mini.greater + run_choice ~ run , value.var='pct')) +
geom_point(
aes(x=mini, y=nova, color=run_choice)) +
facet_wrap(~reorder(interaction(sample.num,mini.greater), sample.num %in% repooled.samples)) +
scale_x_log10() +
scale_y_log10() +
geom_abline() +
labs(x='Miniseq Read Fraction', y='Novaseq Read Fraction',
color='Run choice based on # CAR dropouts (<6)',
shape='sign(Novaseq - Miniseq) read count',
size='log10(abs(Novaseq - Miniseq)) read count')## Warning: Transformation introduced infinite values in continuous x-axis## Warning: Transformation introduced infinite values in continuous y-axis## Warning: Removed 41 rows containing missing values (geom_point).
# compare to baseline
ggplot(
dcast(cyto.read.counts.combined[, `:=`(log_counts= log10(counts), pct= counts / sum(counts)), by=c('run','sample.num')][, pct_v_baseline := pct / pct[bin=='baseline'], by=.(CAR.align, run, donor, t.type, k.type)], CAR.align + sample.num ~ run , value.var='pct_v_baseline')) +
geom_point(
aes(x=mini, y=nova, color=sample.num %in% repooled.samples)) +
facet_wrap(~reorder(sample.num, sample.num %in% repooled.samples)) +
scale_x_log10() +
scale_y_log10() +
geom_abline() +
labs(x='Miniseq Read Fraction', y='Novaseq Read Fraction',
color='Repooled Sample', shape='sign(Novaseq - Miniseq) read count',
size='log10(abs(Novaseq - Miniseq)) read count')## Warning: Transformation introduced infinite values in continuous x-axis## Warning: Transformation introduced infinite values in continuous y-axis## Warning: Removed 166 rows containing missing values (geom_point).
# compare number of cells
ggplot(cyto.read.counts.nova[sort.group %in% c('67_il2.ifng_CD8_pos', '68_il2.ifng_CD8_pos')]) + geom_col(aes(y=log10(car.bin.cells), x=CAR.align, fill=(car.bin.pct))) + facet_grid(sort.cond~donor) + geom_hline(aes(yintercept=log10(100)))+ coord_flip() + scale_fill_distiller(palette='Spectral')## Warning: Removed 5 rows containing missing values (geom_col).
# bin cell count vs sort-group cell counts
ggplot(cyto.read.counts.mini[sort.group %in% c('67_il2.ifng_CD8_pos', '68_il2.ifng_CD8_pos')][, bin.v.dneg := car.bin.cells/car.bin.cells[bin=='dneg'], by=.(CAR.align,sort.group,donor)]) + geom_point(aes(x=log10(car.sort.group.cells), y=log10(car.bin.pct), color=len)) + facet_grid(sort.cond~donor, scales='free') + scale_fill_distiller(palette='Spectral') + ggplot(cyto.read.counts.nova[sort.group %in% c('67_il2.ifng_CD8_pos', '68_il2.ifng_CD8_pos')][, bin.v.dneg := car.bin.cells/car.bin.cells[bin=='dneg'], by=.(CAR.align,sort.group,donor)]) + geom_point(aes(x=log10(car.sort.group.cells), y=log10(car.bin.pct), color=len)) + facet_grid(sort.cond~donor, scales='free') + scale_fill_distiller(palette='Spectral') + geom_abline()
cyto.read.counts.summed <- add_dpos_incl(calc_bin_stats(
cyto.read.counts.combined[
run_choice == run | (run_choice == 'both' & run == 'mini')][,
counts := counts.sum]))
ggplot(cyto.read.counts.mini[sort.group %in% c('67_il2.ifng_CD8_pos', '68_il2.ifng_CD8_pos')][, bin.v.dneg := car.bin.cells/car.bin.cells[bin=='dneg'], by=.(CAR.align,sort.group,donor)]) + geom_point(aes(x=log10(car.sort.group.cells), y=log10(car.bin.pct), color=len)) + facet_grid(sort.cond~donor, scales='free') + scale_fill_distiller(palette='Spectral') + ggplot(cyto.read.counts.nova[sort.group %in% c('67_il2.ifng_CD8_pos', '68_il2.ifng_CD8_pos')][, bin.v.dneg := car.bin.cells/car.bin.cells[bin=='dneg'], by=.(CAR.align,sort.group,donor)]) + geom_point(aes(x=log10(car.sort.group.cells), y=log10(car.bin.pct), color=len)) + facet_grid(sort.cond~donor, scales='free') + scale_fill_distiller(palette='Spectral') + geom_abline() + ggplot(cyto.read.counts.summed[sort.group %in% c('67_il2.ifng_CD8_pos', '68_il2.ifng_CD8_pos')][, bin.v.dneg := car.bin.cells/car.bin.cells[bin=='dneg'], by=.(CAR.align,sort.group,donor)]) + geom_point(aes(x=log10(car.sort.group.cells), y=log10(car.bin.pct), color=len)) + facet_grid(sort.cond~donor, scales='free') + scale_fill_distiller(palette='Spectral') + geom_abline()
ggplot(cyto.read.counts.summed[sort.group %in% c('67_il2.ifng_CD8_pos', '68_il2.ifng_CD8_pos')][, bin.v.dneg := car.bin.cells/car.bin.cells[bin=='dneg'], by=.(CAR.align,sort.group,donor)], aes(x=log10(car.sort.group.cells), y=log10(car.bin.pct), color=len, label=CAR.align)) + geom_point() + facet_grid(sort.cond~donor, scales='free') + scale_fill_distiller(palette='Spectral') + geom_vline(xintercept=log10(500)) + geom_text_repel()## Warning: ggrepel: 36 unlabeled data points (too many overlaps). Consider
## increasing max.overlaps## Warning: ggrepel: 34 unlabeled data points (too many overlaps). Consider
## increasing max.overlaps## Warning: ggrepel: 32 unlabeled data points (too many overlaps). Consider
## increasing max.overlaps## Warning: ggrepel: 29 unlabeled data points (too many overlaps). Consider
## increasing max.overlaps## Warning: ggrepel: 34 unlabeled data points (too many overlaps). Consider
## increasing max.overlaps## Warning: ggrepel: 29 unlabeled data points (too many overlaps). Consider
## increasing max.overlaps## Warning: ggrepel: 32 unlabeled data points (too many overlaps). Consider
## increasing max.overlaps## Warning: ggrepel: 34 unlabeled data points (too many overlaps). Consider
## increasing max.overlaps
Plot Bin Pcts
dcast(rbind(cyto.read.counts.nova[, run_group := 'nova'], cyto.read.counts.mini[, run_group := 'mini'])[sort.reads > min.sort.reads & CAR.align != 'BTLA' &
car.sort.group.cells > 500][,
`:=`(
min_counts= min(counts, na.rm=T),
min_cells= min(car.sort.group.cells, na.rm=T)),
by=c('k.type','assay','t.type','CAR.align','bin','run_group')],
k.type+assay+t.type+bin+CAR.align+min_counts+min_cells+donor~run_group, value.var='car.bin.pct.incl')## k.type assay t.type bin CAR.align min_counts min_cells donor
## 1: neg baseline CD4 baseline 4-1BB 1271 22587.454 67
## 2: neg baseline CD4 baseline 4-1BB 1271 22587.454 68
## 3: neg baseline CD4 baseline 4-1BB 11670 7704.276 67
## 4: neg baseline CD4 baseline 4-1BB 11670 7704.276 68
## 5: neg baseline CD4 baseline BAFF-R 375 6664.276 67
## ---
## 2749: pos tnfa CD8 tnfa TACI 498 1557.763 68
## 2750: pos tnfa CD8 tnfa TIGIT 95 2243.433 67
## 2751: pos tnfa CD8 tnfa TIGIT 95 2243.433 68
## 2752: pos tnfa CD8 tnfa TIGIT 863 2218.227 67
## 2753: pos tnfa CD8 tnfa TIGIT 863 2218.227 68
## mini nova
## 1: 1.0000000 NA
## 2: 1.0000000 NA
## 3: NA 1.0000000
## 4: NA 1.0000000
## 5: 1.0000000 NA
## ---
## 2749: 0.3715896 NA
## 2750: NA 0.4385028
## 2751: NA 0.5043892
## 2752: 0.3910281 NA
## 2753: 0.7192414 NAindiv_run_plots <- function(cyto.read.counts) {
plotlist = list()
plotlist[['car.by.reads']] <- ggplot(cyto.read.counts) +
geom_point(aes(
x=reorder(CAR.align, sort.reads),
y=log10(sort.reads),
color=donor,
shape=sort.reads < min.sort.reads)) +
facet_wrap(~donor+assay+t.type+k.type, nrow=2) +
geom_hline(yintercept=log10(min.sort.reads), linetype=2) +
coord_flip()
plotlist[['car.by.abund']] <- ggplot(
unique(cyto.read.counts[sort.reads > min.sort.reads,
mean.car.abund, by=CAR.align])) +
geom_bar(
aes(
y=mean.car.abund,
x=reorder(CAR.align,mean.car.abund, na.rm=T)),
stat='identity') +
coord_flip() +
labs(title='Mean CAR abundances as fraction of sort')
plotlist[['bin.pcts']] <- ggplot(
cyto.read.counts[!(bin %in% c('neg','baseline','dneg')) &
sort.reads > min.sort.reads]) +
geom_point(
aes(y=car.bin.pct, x=reorder(CAR.align,sort.reads), color=bin)) +
facet_grid(k.type~assay+t.type+donor, scales='free') +
coord_flip()
# compare donors
compare_donors <- dcast(
cyto.read.counts[!(bin %in% c('neg','baseline','dneg')) &
sort.reads > min.sort.reads & CAR.align != 'BTLA' &
car.sort.group.cells > 500][,
`:=`(
min_counts= min(counts, na.rm=T),
min_cells= min(car.sort.group.cells, na.rm=T)),
by=c('k.type','assay','t.type','CAR.align','bin')],
k.type+assay+t.type+bin+CAR.align+min_counts+min_cells~donor,
value.var='car.bin.pct.incl')
compare_donors <- compare_donors[,
`:=`(top_67= rank(-`67`) < 5,
top_68= rank(-`68`) < 5,
favorites= CAR.align %in% c(
'BAFF-R','NTB-A','TACI','4-1BB','CD28','CD40','CD2')),
by=.(k.type, assay, t.type, bin)]
compare_donors[, label := ifelse(top_67 | top_68 | favorites, CAR.align, '')]
plotlist[['bin.pcts.donor.comp']] <- ggplot(
compare_donors, aes(x=`67`, y=`68`, color=log10(min_cells), label=label)) +
geom_point() +
facet_wrap(k.type~assay+t.type+bin, scales='free') +
geom_text_repel() +
scale_x_continuous(labels=scales::percent) +
scale_y_continuous(labels=scales::percent)
plotlist[['bin.pcts.donor.comp.il2']] <- ggplot(
compare_donors[bin=='il2'], aes(x=`67`, y=`68`, color=bin, label=label)) +
geom_point() +
facet_wrap(k.type~assay+t.type+bin, scales='free') +
geom_text_repel() +
scale_x_continuous(labels=scales::percent) +
scale_y_continuous(labels=scales::percent)
plotlist[['bin.pcts.donor.comp.il2.ifng']] <- ggplot(
compare_donors[assay=='il2.ifng'][,
`:=`(top_67= rank(-`67`) < 10,
top_68= rank(-`68`) < 10,
favorites= CAR.align %in% c(
'BAFF-R','NTB-A','TACI','4-1BB','CD28','CD40','CD2')),
by=.(k.type, assay, t.type, bin)][,
label := ifelse(top_67 | top_68 | favorites, CAR.align, '')],
aes(x=`67`, y=`68`, color=bin, label=label)) +
geom_point() +
facet_wrap(k.type~assay+t.type+bin, scales='free') +
geom_text_repel() +
scale_x_continuous(labels=scales::percent) +
scale_y_continuous(labels=scales::percent)
return(plotlist)
}
#compare runs (mini vs nova)
cyto.read.counts.runcomp <- dcast(
rbind(cyto.read.counts.nova[, run_group := 'nova'],
cyto.read.counts.mini[, run_group := 'mini'])[
sort.reads > min.sort.reads & CAR.align != 'BTLA' &
!(bin %in% c('neg','baseline','dneg')) &
car.sort.group.cells > 500][,
`:=`(
min_counts= min(counts, na.rm=T),
min_cells= min(car.sort.group.cells, na.rm=T)),
by=c('k.type','assay','t.type','CAR.align','bin','donor')],
k.type+assay+t.type+bin+CAR.align+donor+min_counts+min_cells~run_group,
value.var='car.bin.pct')
cyto.read.counts.runcomp <- cyto.read.counts.runcomp[,
`:=`(top_mini= rank(-`mini`) < 5,
top_nova= rank(-`nova`) < 5,
favorites= CAR.align %in% c(
'BAFF-R','NTB-A','TACI','4-1BB','CD28','CD40','CD2')),
by=.(k.type, assay, t.type, bin, donor)][,
label := ifelse(top_mini | top_nova | favorites, CAR.align, '')]
# ggplot(
# cyto.read.counts.runcomp, aes(x=`mini`, y=`nova`, label=label)) +
# geom_point() +
# facet_wrap(k.type~assay+t.type+bin+donor, scales='free') +
# geom_text_repel() +
# scale_x_continuous(labels=scales::percent) +
# scale_y_continuous(labels=scales::percent)Plot Bin Pcts - combined read counts
indiv_run_plots <- function(cyto.read.counts) {
plotlist = list()
plotlist[['car.by.reads']] <- ggplot(cyto.read.counts) +
geom_point(aes(
x=reorder(CAR.align, sort.reads),
y=log10(sort.reads),
color=donor,
shape=sort.reads < min.sort.reads)) +
facet_wrap(~donor+assay+t.type+k.type, nrow=2) +
geom_hline(yintercept=log10(min.sort.reads), linetype=2) +
coord_flip()
plotlist[['car.by.abund']] <- ggplot(
unique(cyto.read.counts[sort.reads > min.sort.reads,
mean.car.abund, by=CAR.align])) +
geom_bar(
aes(
y=mean.car.abund,
x=reorder(CAR.align,mean.car.abund, na.rm=T)),
stat='identity') +
coord_flip() +
labs(title='Mean CAR abundances as fraction of sort')
plotlist[['bin.pcts']] <- ggplot(
cyto.read.counts[!(bin %in% c('neg','baseline','dneg')) &
sort.reads > min.sort.reads]) +
geom_point(
aes(y=car.bin.pct, x=reorder(CAR.align,sort.reads), color=bin)) +
facet_grid(k.type~assay+t.type+donor, scales='free') +
coord_flip()
# compare donors
compare_donors <- dcast(
cyto.read.counts[!(bin %in% c('neg','baseline','dneg')) &
sort.reads > min.sort.reads &
car.sort.group.cells > 500][,
`:=`(
min_counts= min(counts, na.rm=T),
min_cells= min(car.sort.group.cells, na.rm=T)),
by=c('k.type','assay','t.type','CAR.align','bin')],
k.type+assay+t.type+bin+CAR.align+min_counts+min_cells~donor,
value.var='car.bin.pct.incl')
compare_donors <- compare_donors[,
`:=`(top_67= rank(-`67`) < 5,
top_68= rank(-`68`) < 5,
favorites= CAR.align %in% c(
'BAFF-R','NTB-A','TACI','4-1BB','CD28','CD40','CD2')),
by=.(k.type, assay, t.type, bin)]
compare_donors[, label := ifelse(top_67 | top_68 | favorites, CAR.align, '')]
plotlist[['bin.pcts.donor.comp']] <- ggplot(
compare_donors, aes(x=`67`, y=`68`, color=min_cells, label=label)) +
geom_point() +
facet_wrap(k.type~assay+t.type+bin, scales='free') +
geom_text_repel() +
scale_x_continuous(labels=scales::percent) +
scale_y_continuous(labels=scales::percent)
plotlist[['bin.pcts.donor.comp.il2']] <- ggplot(
compare_donors[bin=='il2'], aes(x=`67`, y=`68`, color=bin, label=label)) +
geom_point() +
facet_wrap(k.type~assay+t.type+bin, scales='free') +
geom_text_repel() +
scale_x_continuous(labels=scales::percent) +
scale_y_continuous(labels=scales::percent)
plotlist[['bin.pcts.donor.comp.il2.ifng']] <- ggplot(
compare_donors[assay=='il2.ifng'][,
`:=`(top_67= rank(-`67`) < 10,
top_68= rank(-`68`) < 10,
favorites= CAR.align %in% c(
'BAFF-R','NTB-A','TACI','4-1BB','CD28','CD40','CD2')),
by=.(k.type, assay, t.type, bin)][,
label := ifelse(top_67 | top_68 | favorites, CAR.align, '')],
aes(x=`67`, y=`68`, color=bin, label=label)) +
geom_point() +
facet_wrap(k.type~assay+t.type+bin, scales='free') +
geom_text_repel() +
scale_x_continuous(labels=scales::percent) +
scale_y_continuous(labels=scales::percent)
return(plotlist)
}
#compare runs (mini vs nova)
cyto.read.counts.runcomp <- dcast(
rbind(cyto.read.counts.nova[, run_group := 'nova'],
cyto.read.counts.mini[, run_group := 'mini'])[
sort.reads > min.sort.reads & CAR.align != 'BTLA' &
!(bin %in% c('neg','baseline','dneg')) &
car.sort.group.cells > 500][,
`:=`(
min_counts= min(counts, na.rm=T),
min_cells= min(car.sort.group.cells, na.rm=T)),
by=c('k.type','assay','t.type','CAR.align','bin','donor')],
k.type+assay+t.type+bin+CAR.align+donor+min_counts+min_cells~run_group,
value.var='car.bin.pct')
cyto.read.counts.runcomp <- cyto.read.counts.runcomp[,
`:=`(top_mini= rank(-`mini`) < 5,
top_nova= rank(-`nova`) < 5,
favorites= CAR.align %in% c(
'BAFF-R','NTB-A','TACI','4-1BB','CD28','CD40','CD2')),
by=.(k.type, assay, t.type, bin, donor)][,
label := ifelse(top_mini | top_nova | favorites, CAR.align, '')]
# ggplot(
# cyto.read.counts.runcomp, aes(x=`mini`, y=`nova`, label=label)) +
# geom_point() +
# facet_wrap(k.type~assay+t.type+bin+donor, scales='free') +
# geom_text_repel() +
# scale_x_continuous(labels=scales::percent) +
# scale_y_continuous(labels=scales::percent)
run_plots <- indiv_run_plots(
cyto.read.counts.summed[
sort.reads > min.sort.reads & car.sort.group.cells > 500][,
`:=`(
min_counts= min(counts, na.rm=T),
min_cells= min(car.sort.group.cells, na.rm=T))])
compare_48 <- dcast(
cyto.read.counts.summed[!(bin %in% c('neg','baseline','dneg')) &
sort.reads > min.sort.reads &
car.sort.group.cells > 500][,
`:=`(
min_counts= min(counts, na.rm=T),
min_cells= min(car.sort.group.cells, na.rm=T)),
by=c('k.type','assay','donor','CAR.align','bin')],
k.type+assay+donor+bin+CAR.align+min_counts+min_cells~t.type,
value.var='car.bin.pct.incl')
compare_48 <- compare_48[,
`:=`(top_4= rank(-CD4) < 5,
top_8= rank(-CD8) < 5,
favorites= CAR.align %in% c(
'BAFF-R','NTB-A','TACI','4-1BB','CD28','CD40','CD2')),
by=.(k.type, assay, donor, bin)]
compare_48[, label := ifelse(top_8 | top_4 | favorites, CAR.align, '')]
ggplot(
compare_48[!is.na(CD4) & !is.na(CD8)],
aes(x=CD4, y=CD8, color=bin, label=label)) +
geom_point() +
facet_wrap(k.type~assay+donor+bin, scales='free') +
geom_text_repel() +
scale_x_continuous(labels=scales::percent) +
scale_y_continuous(labels=scales::percent)## Warning: ggrepel: 5 unlabeled data points (too many overlaps). Consider
## increasing max.overlaps## Warning: ggrepel: 3 unlabeled data points (too many overlaps). Consider
## increasing max.overlaps## Warning: ggrepel: 4 unlabeled data points (too many overlaps). Consider
## increasing max.overlaps## Warning: ggrepel: 2 unlabeled data points (too many overlaps). Consider increasing max.overlaps
## ggrepel: 2 unlabeled data points (too many overlaps). Consider increasing max.overlaps
## ggrepel: 2 unlabeled data points (too many overlaps). Consider increasing max.overlaps## Warning: ggrepel: 1 unlabeled data points (too many overlaps). Consider
## increasing max.overlaps## Warning: ggrepel: 2 unlabeled data points (too many overlaps). Consider
## increasing max.overlaps## Warning: ggrepel: 4 unlabeled data points (too many overlaps). Consider
## increasing max.overlaps## Warning: ggrepel: 2 unlabeled data points (too many overlaps). Consider
## increasing max.overlaps
# ggplot(dcast(cyto.read.counts.summed[sort.group %in% c("67_baseline_CD8_neg","67_baseline_CD8_pos")], assay+t.type+bin+CAR.align~k.type, value.var='car.abund'), aes(x=log10(neg), y=log10(pos), label=CAR.align)) + geom_point() + geom_text_repel() + geom_abline()
#
# ggplot(dcast(cyto.read.counts.summed[sort.group %in% c("68_baseline_CD8_neg","68_baseline_CD8_pos")], assay+t.type+bin+CAR.align~k.type, value.var='car.abund'), aes(x=log10(neg), y=log10(pos), label=CAR.align)) + geom_point() + geom_text_repel() + geom_abline()
#
# ggplot(dcast(cyto.read.counts.summed[sort.group %in% c("68_baseline_CD8_neg","67_baseline_CD8_neg")], assay+t.type+bin+CAR.align~donor, value.var='car.abund'), aes(x=log10(`67`), y=log10(`68`), label=CAR.align)) + geom_point() + geom_text_repel() + geom_abline()
#
# ggplot(dcast(cyto.read.counts.summed[sort.group %in% c("68_baseline_CD8_pos","67_baseline_CD8_pos")], assay+t.type+bin+CAR.align~donor, value.var='car.abund'), aes(x=log10(`67`), y=log10(`68`), label=CAR.align)) + geom_point() + geom_text_repel() + geom_abline()Combine with Old Donor
old <- new.env()
load(file.path(data.dir, 'pooled_read_counts.Rdata'), envir=old)
old.read.counts <- old$read.counts[assay %in% c('IL2','IFNy')]
old.read.counts[assay == 'IFNy', assay := 'ifng']
old.read.counts[assay == 'IL2', assay := 'il2']
old.read.counts[assay == 'ifng', sort.cond := ifelse(bin == 'A', 'neg','ifng')]
old.read.counts[assay == 'il2', sort.cond := ifelse(bin == 'A', 'neg','il2')]
old.read.counts <- old.read.counts[, list(
CAR.align,
cell.count,
bin.pct,
car.bin.pct,
car.bin.cells= cell.count*car.bin.read.pct),
by=.(donor, assay, bin, sort.cond, k.type, t.type)]
old.read.counts <- old.read.counts[, list(
cell.count= cell.count[1],
car.bin.cells = sum(car.bin.cells),
bin.pct= bin.pct[1],
car.bin.pct.incl= sum(car.bin.pct),
car.bin.pct= sum(car.bin.pct)),
by=.(sort.cond, CAR.align, donor, assay, t.type, k.type)]
old.read.counts[, sort.group := paste(donor, assay, t.type, k.type, sep='_')]
old.read.counts[, car.sort.group.cells := sum(car.bin.cells),
by=.(CAR.align, sort.group)]
shared_cols <- intersect(
names(old.read.counts),
names(cyto.read.counts.summed))
cyto.read.counts.all <- rbind(
cyto.read.counts.summed[, ..shared_cols],
old.read.counts[, ..shared_cols])IL2 & IFNG plots with old donor added
il2.ifng.plots <- list()
# IFNG and IL2 donor vs donor each (the lazy ugly way)
il2.ifng.plots$between.donors <- ((ggplot(dcast(cyto.read.counts.all[k.type == 'pos' & t.type == 'CD4' & assay %in% c('il2.ifng', 'ifng') & car.sort.group.cells > 500], sort.cond+CAR.align+t.type+k.type~donor, value.var = 'car.bin.pct.incl')[sort.cond == 'ifng'], aes(x=`67`, y=`d1`)) + geom_point() + geom_text_repel(aes(label=CAR.align))) + (ggplot(dcast(cyto.read.counts.all[k.type == 'pos' & t.type == 'CD4' & assay %in% c('il2.ifng', 'ifng') & car.sort.group.cells > 500], sort.cond+CAR.align+t.type+k.type~donor, value.var = 'car.bin.pct.incl')[sort.cond == 'ifng'], aes(x=`68`, y=`d1`)) + geom_point() + geom_text_repel(aes(label=CAR.align))) + (ggplot(dcast(cyto.read.counts.all[k.type == 'pos' & t.type == 'CD4' & assay %in% c('il2.ifng', 'ifng') & car.sort.group.cells > 500], sort.cond+CAR.align+t.type+k.type~donor, value.var = 'car.bin.pct.incl')[sort.cond == 'ifng'], aes(x=`68`, y=`67`)) + geom_point() + geom_text_repel(aes(label=CAR.align))) + plot_annotation(
title = 'IFNy Donor vs Donor')) /
((ggplot(dcast(cyto.read.counts.all[k.type == 'pos' & t.type == 'CD4' & assay %in% c('il2.ifng', 'il2') & car.sort.group.cells > 500], sort.cond+CAR.align+t.type+k.type~donor, value.var = 'car.bin.pct.incl')[sort.cond == 'il2'], aes(x=`67`, y=`d1`)) + geom_point() + geom_text_repel(aes(label=CAR.align))) + (ggplot(dcast(cyto.read.counts.all[k.type == 'pos' & t.type == 'CD4' & assay %in% c('il2.ifng', 'il2') & car.sort.group.cells > 500], sort.cond+CAR.align+t.type+k.type~donor, value.var = 'car.bin.pct.incl')[sort.cond == 'il2'], aes(x=`68`, y=`d1`)) + geom_point() + geom_text_repel(aes(label=CAR.align))) + (ggplot(dcast(cyto.read.counts.all[k.type == 'pos' & t.type == 'CD4' & assay %in% c('il2.ifng', 'il2') & car.sort.group.cells > 500], sort.cond+CAR.align+t.type+k.type~donor, value.var = 'car.bin.pct.incl')[sort.cond == 'il2'], aes(x=`68`, y=`67`)) + geom_point() + geom_text_repel(aes(label=CAR.align))) + plot_annotation(
title = 'IL2 Donor vs Donor'))
#within donor, 4v8
il2.ifng.plots$within.donor.4v8 <- ggplot(dcast(cyto.read.counts.all[k.type == 'pos' & assay %in% c('il2.ifng', 'ifng','il2') & car.sort.group.cells > 500 & sort.cond %in% c('ifng','il2')], donor+CAR.align+k.type+sort.cond~t.type, value.var = 'car.bin.pct.incl')[], aes(x=`CD4`, y=`CD8`)) + geom_point() + geom_text_repel(aes(label=CAR.align)) + facet_wrap(sort.cond~donor, scales='free')
#within donor, il2.v.ifng
il2.ifng.plots$within.donor.il2vifng <- ggplot(dcast(cyto.read.counts.all[k.type == 'pos' & assay %in% c('il2.ifng', 'ifng','il2') & car.sort.group.cells > 500 & sort.cond %in% c('ifng','il2')], donor+CAR.align+t.type+k.type~sort.cond, value.var = 'car.bin.pct.incl')[], aes(x=`ifng`, y=`il2`)) + geom_point() + geom_text_repel(aes(label=CAR.align)) + facet_wrap(t.type~donor, scales='free')
cyto.ranked.all <- cyto.read.counts.all[
k.type == 'pos' & assay %in% c('il2.ifng', 'il2','ifng') &
sort.cond %in% c('il2','ifng') &
car.sort.group.cells > 400]
#individual ifng/il2
ggplot(cyto.ranked.all[,
`:=`(
sort.group.cond.rank= rank(-car.bin.pct.incl),
CAR.align.sort.cond= factor(
paste(CAR.align,
paste(sort.cond, sep='_'), sep='|'))),
by=.(sort.cond, sort.group)]) +
geom_point(
aes(x=reorder(CAR.align.sort.cond,-sort.group.cond.rank),
y=sort.group.cond.rank, color=donor, shape=t.type),
position='dodge') +
scale_x_discrete(label=function(str) gsub('\\|.*','',str)) +
facet_wrap( ~ sort.cond, scales='free') + coord_flip()## Warning: Width not defined. Set with `position_dodge(width = ?)`
#cd4/cd8 separate combined ifng/il2
ggplot(cyto.ranked.all[,
`:=`(
sort.group.cond.rank= rank(-car.bin.pct.incl),
CAR.align.sort.cond= factor(
paste(CAR.align,
paste(t.type, sep='_'), sep='|'))),
by=.(sort.cond, sort.group)]) +
geom_point(
aes(x=reorder(CAR.align.sort.cond,-sort.group.cond.rank),
y=sort.group.cond.rank, color=donor, shape=sort.cond),
position='dodge') +
scale_x_discrete(label=function(str) gsub('\\|.*','',str)) +
facet_wrap(t.type ~ sort.cond, scales='free') + coord_flip()## Warning: Width not defined. Set with `position_dodge(width = ?)`
#cd4/cd8 combined ifng/il2
ggplot(cyto.ranked.all[,
`:=`(
sort.group.cond.rank= rank(-car.bin.pct.incl),
CAR.align.sort.cond= factor(
paste(CAR.align,
paste(t.type, sep='_'), sep='|'))),
by=.(sort.cond, sort.group)]) +
geom_point(
aes(x=reorder(CAR.align.sort.cond,-sort.group.cond.rank),
y=sort.group.cond.rank, color=donor, shape=sort.cond),
position='dodge') +
scale_x_discrete(label=function(str) gsub('\\|.*','',str)) +
facet_wrap(t.type ~ ., scales='free') + coord_flip()## Warning: Width not defined. Set with `position_dodge(width = ?)`
#cd4/cd8 combined ifng/il2 separate donor
ggplot(cyto.ranked.all[,
`:=`(
sort.group.cond.rank= rank(-car.bin.pct.incl),
CAR.align.sort.cond= factor(
paste(CAR.align,
paste(sort.cond, t.type, sep='_'), sep='|'))),
by=.(sort.cond, sort.group)]) +
geom_point(
aes(x=reorder(CAR.align.sort.cond,-sort.group.cond.rank),
y=sort.group.cond.rank, color=donor, shape=t.type),
position='dodge') +
scale_x_discrete(label=function(str) gsub('\\|.*','',str)) +
facet_wrap(t.type ~ sort.cond, scales='free') + coord_flip()## Warning: Width not defined. Set with `position_dodge(width = ?)`
#all combined ifng/il2
ggplot(cyto.ranked.all[car.sort.group.cells > 600][,
`:=`(
sort.group.cond.rank= rank(-car.bin.pct.incl),
CAR.align.sort.cond= factor(CAR.align)),
by=.(sort.cond, sort.group)]) +
geom_jitter(height=0,
aes(x=reorder(CAR.align.sort.cond,-sort.group.cond.rank),
y=sort.group.cond.rank, color=donor, shape=interaction(t.type, sort.cond))) +
coord_flip()
#total cytokine secretion
ggplot(cyto.ranked.all[car.sort.group.cells > 600][,
`:=`(
sort.group.cond.rank= rank(-car.bin.pct.incl),
CAR.align.sort.cond= factor(CAR.align)),
by=.(sort.cond, sort.group)]) +
geom_boxplot(
aes(x=reorder(CAR.align.sort.cond,-sort.group.cond.rank),
y=sort.group.cond.rank), outlier.size = NULL) +
coord_flip() +
labs(
title='CARS ranked by overall IFNg/IL2 secretion',
subtitle='3 Donors for CD4, 2 Donors for CD8; at least 600 cells per sort') +
theme(panel.grid=element_blank())
### Two Sample Wilcoxon Tests with Benjamini-Hochberg Correction
library(rstatix)##
## Attaching package: 'rstatix'## The following object is masked from 'package:stats':
##
## filterdo_wilcox_test <- function(df) {
wilcox_test(df,
sort.group.cond.rank ~ CAR.align,
ref.group='all',
alternative='g') %>%
adjust_pvalue(method = "BH") %>%
add_significance("p.adj")
}
car_ranks <- cyto.ranked.all[car.sort.group.cells > 500][,
`:=`(
sort.group.cond.rank= rank(-car.bin.pct.incl),
CAR.align.sort.cond= factor(CAR.align)),
by=.(sort.cond, sort.group)]
wilcoxon_stats <- rbind(
data.table(do_wilcox_test(car_ranks))[, cond := 'All Conditions'][, group := 'all'],
car_ranks[,
do_wilcox_test(.SD),
by='sort.cond'][, group := 'Per Cytokine'] %>%
setnames(., 'sort.cond', 'cond'),
car_ranks[,
do_wilcox_test(.SD),
by='donor'][, group := 'Per Donor'] %>%
setnames(., 'donor', 'cond'),
car_ranks[,
do_wilcox_test(.SD),
by='t.type'][, group := 'Per T Cell Subset'] %>%
setnames(., 't.type', 'cond'))
significance_rankings <- ggplot(wilcoxon_stats) +
geom_tile(aes(x=reorder(group2,-log(p.adj)), y=cond, fill=-log(p.adj)), size=2) +
geom_tile(
data=wilcoxon_stats[p.adj.signif != 'ns'],
aes(x=reorder(group2,-log(p.adj)), y=cond),
color='white', fill=NA, size=2) +
coord_flip() +
scale_fill_distiller(palette='Oranges', direction=1) + theme_minimal() +
theme(panel.grid=element_blank(), panel.spacing = unit(0, "lines")) +
facet_grid(~group, scales='free') +
labs(
title= 'CAR cytokine secretion rankings and significance',
subtitle='Mann-Whitney w/ Benjamini-Hochberg correction (one-sided)',
caption='White border: adjusted p-value < 0.05. Missing tile: Removed CARs with <500 cells/sort.',
fill='-log(p)',
x='Costimulatory domain',
y='Cytokine Condition')
t.type_ranks <- dcast(car_ranks, CAR.align+donor+sort.cond ~ t.type,
value.var='sort.group.cond.rank')
donor_ranks <- dcast(car_ranks, CAR.align+t.type+sort.cond ~ donor,
value.var='sort.group.cond.rank')
sort.cond_ranks <- dcast(car_ranks, CAR.align+t.type+donor ~ sort.cond,
value.var='sort.group.cond.rank')
cor(na.omit(donor_ranks[, list(`67`, `68`, `d1`)]), method='spearman')## 67 68 d1
## 67 1.0000000 0.5635049 0.5068793
## 68 0.5635049 1.0000000 0.6412498
## d1 0.5068793 0.6412498 1.0000000cor(na.omit(t.type_ranks[, list(CD4, CD8)]), method='spearman')## CD4 CD8
## CD4 1.0000000 0.7017806
## CD8 0.7017806 1.0000000cor(na.omit(sort.cond_ranks[, list(ifng, il2)]), method='spearman')## ifng il2
## ifng 1.0000000 0.7021237
## il2 0.7021237 1.0000000data.output.dir <- file.path(s3.data.dir)
save(list=c('cyto.read.counts.summed'),
file=file.path(data.output.dir, 'pooled_cytokine_repeat_data.Rdata'))